Article Text

Download PDFPDF

Fine-tuning the impairment argument
  1. Bruce Philip Blackshaw1,
  2. Perry Hendricks2
  1. 1 Philosophy, University of Birmingham, Birmingham, UK
  2. 2 Department of Philosophy, Purdue University, West Lafayette, Indiana, USA
  1. Correspondence to Bruce Philip Blackshaw, Philosophy, University of Birmingham, Birmingham B15 2TT, UK; bblackshaw{at}


Perry Hendricks’ original impairment argument for the immorality of abortion is based on the impairment principle: if impairing an organism to some degree is immoral, then ceteris paribus, impairing it to a higher degree is also immoral. Since abortion impairs a fetus to a higher degree than fetal alcohol syndrome (FAS) and giving a fetus FAS is immoral, it follows that abortion is immoral. Critics have argued that the ceteris paribus is not met for FAS and abortion, and so we proposed the modified impairment principle (MIP) to avoid these difficulties. Dustin Crummett has responded, arguing that MIP is open to various counterexamples which show it to be false. He also shows that MIP can generate moral dilemmas. Here, we propose a modification to MIP that resolves the issues Crummett raises. Additionally, Alex Gillham has criticised our appropriation of Don Marquis’ ‘future like ours’ reasoning about the wrongness of impairment. We show that his objections have minimal implications for our argument.

  • abortion
  • moral status
  • embryos and fetuses
  • ethics

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Perry Hendricks’ impairment argument claims that since it is immoral to impair a fetus by giving it fetal alcohol syndrome (FAS), it is ceteris paribus immoral to have an abortion, since that impairs the fetus to a higher degree. This follows from his impairment principle, which states that if it is immoral to impair an organism to the nth degree, then ceteris paribus it is immoral to impair it to the n+1 degree.1 2 We attempted to strengthen the argument by replacing the ceteris paribus clause with the modified impairment principle (MIP).3 However, Dustin Crummett has argued that MIP is false4 and Alex Gillham has criticised our appropriation of Don Marquis’ ‘future like ours’ (FLO) reasoning.5 Here, we respond to these criticisms.


Crummett argues that MIP faces counterexamples whenever the n+1 degree impairment can be justified by a sufficiently good reason that is not present for an nth degree impairment.4 He provides an example of flooding a room you are in with gas. In scenario (A), done for no reason, this hinders your ability to roll your tongue. In (B), this removes your ability to roll your tongue. However, in (B), this is done because otherwise the gas would flood into another room with people allergic to it, killing them.

Clearly, in (B) releasing the gas is permissible, and so this is a counterexample to MIP. As Crummett notes, Hendricks’ ceteris paribus clause prevented such counterexamples, since a valuable good obtains in one case (saving lives) and not in the other case, meaning that the clause is not satisfied. However, MIP contains no ceteris paribus clause, and so we cannot avoid this result. Hence, we must modify MIP.

Our solution is to add an over-rider provision: a reason R only renders an impairment immoral if it is not over-ridden by an opposing reason R*. To illustrate this, suppose that a reason R makes impairing O to the nth degree immoral. Now, suppose that a reason R* is stronger than R and counts in favour of impairing O to the n+1 degree. If both R and R* obtain in this case, then R is over-ridden and it is not immoral to impair O to the n+1 degree. Put differently, since R* is stronger than R, it over-rides it and renders impairing O to the n+1 degree permissible. (This is a common condition in deontological ethics.6) Crummett’s cases (A) and (B) are a concrete example: in (A) the impairment is immoral, but in (B) the reason impairment would be immoral is over-ridden by an opposing reason (the impairment is necessary to save several lives). The MIP2 may be restated as follows:

MIP2: If it is immoral to impair an organism O to the nth degree for reason R, then, provided R continues to hold (or is present) and there are no over-riding reasons, it is immoral to impair O to the n+1 degree.

When is the over-rider provision satisfied? That is, how do we know when R* over-rides R? If R* over-rides R when impairing O to the n+1 degree, then it will also over-ride R when impairing O to a lesser degree. Take Crummett’s example: if the over-riding reason in case (B) were present in case (A) (ie, if the impairment was needed to save lives), then the impairment referred to in case (A) would be permissible. So, if a reason R* is suggested as an over-riding reason for R, then we can check whether it is one by seeing if it would justify lesser impairments. If it does, then it is an over-riding reason, otherwise, it is not.

Crummett’s example does not meet the over-rider provision: MIP2 does not render it immoral to impair someone’s ability to use her tongue in (B), since saving the lives of several people over-rides our reason for not causing the impairment.

Moral dilemmas

Crummett also argues that MIP can generate genuine moral dilemmas, providing a trolley problem as an example.4 In scenario (C), if I continue driving straight, I will hit someone and sever their leg, but if I turn, nothing bad happens. In scenario (D), if I continue driving straight, I will hit someone and sever both their legs, but if I turn the track loops back then I will hit them anyway.

In case (C) going straight ahead is immoral, since it impairs the person. In (D), going straight is immoral (by MIP, since it is a greater impairment than in (C)), but turning is also immoral. So, both options are immoral—a moral dilemma.

MIP2 does not generate a dilemma. In (D), the impairment is permissible because of an over-riding reason—no matter what the driver does, he will sever the person’s legs. The over-rider provision states that a reason that renders an impairment to the n+1 degree permissible must also render impairment to a lesser degree permissible. Clearly, if the trolley driver in (C) could also not avoid impairing the person, his action would be permissible. Therefore, MIP2 does not apply to (D) because of an over-riding reason, and Crummett’s moral dilemma is avoided.

A future like ours

Gillham objects to the ‘strengthened impairment argument’ (SIA) based on its use of Don Marquis’ FLO as a reason why impairment of a fetus by FAS is immoral.5 He argues that if a fetus will never be born, it has no FLO to be deprived of, and so impairment by FAS would not be immoral.

Before responding to Gillham, it is important to make clear that the impairment argument is not wed to the FLO. We use FLO merely as one possible explanation of the wrongness of impairment.

There are two categories of cases to consider: first, where the fetus is killed unforeseeably, and second, where it is known that the fetus cannot survive to birth. The SIA explicitly refers to the former, and claims that deprivation of FLO provides an explanation of why impairment of a fetus by FAS is immoral, even if a FLO would not have been actual. In such cases, at the time of impairment, the fetus’s FLO is contingent on future events, and this raises the question of what future the fetus is being deprived of. Depending on one’s view of the problem of future contingents, it may not even be true or false that the fetus has a FLO. Mary Clayton Coleman makes a helpful proposal we can adopt—she suggests a ‘for all we know’ (FAWK) revision of Marquis’ account, grounding the wrongness of killing in the deprivation of FLO we believe it will have.7 This highlights the notion of moral responsibility in deprivation of FLOs. Therefore, FAWK FLO explains the wrongness of inflicting FAS in cases of unforeseen death.

Gillham is correct that in cases where we know the fetus cannot survive because of a lethal condition, deprivation of FLO faces difficulties as an explanation. However, even if we grant that SIA fails in such cases, this only applies in a very limited number of cases. The fetus would need to suffer a condition that invariably leads to death before or soon after birth. However, for many so-called ‘lethal malformations’, long-term survival is possible, and so it is not clear that such fetuses lack a FLO.8 Diagnoses of the severity of such conditions also can be inaccurate.9 So granting Gillham’s point is unlikely to affect the applicability of SIA to the vast majority of induced abortions.

Gillham raises one more scenario: FAS followed by a planned induced abortion. The implication is that inflicting FAS is not immoral because at the time of impairment, the fetus has no FLO to be deprived of—we know the abortion will prevent it from arising. There are several possible responses. First, at the time of impairment, we cannot be certain that the abortion will occur—there may be a change of mind, or financial difficulties, or even a pandemic that temporarily shuts down elective medical procedures. Again, the FAWK revision to Marquis can be used here—for all we know the fetus does have a FLO, despite a firm intention to later abort. That is, any event that is a future contingent (accidental or deliberate) does not affect the wrongness of FLO deprivation in the present. Second, we have argued that deprivation of a FLO is a plausible reason for the wrongness of inflicting prenatal injury in cases where the fetus survives and where it is unforeseeably killed—the wrong-making property of inflicting prenatal injury is deprivation of FLO. It is not plausible that we can prevent or alleviate this initial wrong by a further deprivation of FLO. Note that this claim does not require assuming that such further deprivation of FLO makes the abortion itself immoral—that conclusion is derived from SIA.

Alternatively, we could understand the wrongness of prenatal impairment on account of it resulting in the deprivation of a possible FLO. Of course, the fetus has many possible FLOs—so, which one is relevant? The relevant one is the FLO the individual would have enjoyed if she were not impaired (lethally or otherwise); it is the one that would have been actual if she were not impaired. Of course, possible FLOs are contingent on future events, but the fetus nonetheless is deprived of these possible FLOs at the time of the impairment: these possible futures are cut-off from her. The deprivation of possible FLOs (of the relevant sort, described above) provides a plausible explanation of the wrongness of prenatal injury both when the fetus subsequently dies accidentally and when it dies from an action intended to kill it. Whether the fetus is killed (lethally impaired) accidentally or intentionally, it does not change the possible FLO the fetus would have had if it were not to have been killed. And so, this issue is circumvented: impairment is immoral even when one intends to (and does) kill the fetus subsequently.

Restating Marquis?

Finally, Gilliam argues that in cases where SIA succeeds, it merely restates Marquis’ original argument, and so is insignificant.5 However, Marquis makes no reference to prenatal impairment. SIA uses FLO as a plausible reason as to why impairment by FAS is immoral—not abortion. SIA relies on our intuition that this is immoral, not intuitions regarding killing. Second, SIA does not claim that deprivation of FLO, or indeed reason R, is why abortion (or any higher degree of impairment) is immoral. Instead, it says that if R is the reason impairment of the nth degree is wrong, then provided R holds for n+1 degree impairment and the over-rider provision is satisfied, the higher impairment is also wrong. This means that abortion could be wrong for a different reason than why giving a fetus FAS is wrong. So, SIA is importantly different to Marquis’ argument—it is predicated on our intuitions regarding prenatal injury, not a theory of the wrongness of killing.

Ethics statements


Thanks to G.L.G.—Colin Patrick Mitchell—for particularly insightful comments.



  • Contributors BPB wrote the first draft. PH contributed one of the key concepts for this draft—the over-rider provision—and was also responsible for substantial and important revisions. Both authors worked on condensing the original submission to produce the revised submission.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; internally peer reviewed.

Other content recommended for you